Only the fourth paragraph talks about what you just mentioned. From the other paragraphs my takeaway was that startups have to start at a problem, and then happ

Only the fourth paragraph talks about what you just mentioned. From the other paragraphs my takeaway was that startups have to start at a problem, and then happen to make something that can retrospectively be described as x for y, like Edmodo's and SoundCloud's founders did. This as opposed to trying to be x for y just because x or y is hot. I'm not saying the article was good, just pointing out what it said.